The first order help a young person needs is being taken seriously and encouragement in their process of discovery. The important conversation isn't "Use this" in it's strong or weak forms. It's the default StackOverflow comment: "What code have you tried? What error are you getting?" (perhaps in a weaker form, perhaps not, depending on the child).

Let the child own the process and understand that the most likely outcome statistically is that the actual process of creating a computer game will turn out to be unattractive as it is to a first approximation for everyone. Making games is hard for highly intelligent adults - much harder than writing a Rails app.

The mistake that I find easy to make with my own child is an unwillingness to let their interest unfold in its own time as part of the growing process. Adult interests and behaviors and skills take years to develop. The twelve year old boy will be radically different intellectually in two years...or even one. It takes patience and a long-term view and an understanding that many of a child's interests are passing. Some come back, most don't, and what tends to come back are interests that they find themselves sharing with friends. To put it another way, your nephew's English comprehension could be orders of magnitude better in five years. His understanding of mathematics most certainly will.

For concrete advice:

Provide high quality resources - the sort of durable tools and books and websites that one would give to an adult. The interest that wanes in a month at twelve may be rekindled for a year at thirteen and arise from hibernation to become a career choice at twenty four...and there sitting on the bookshelf is a weaker form of The Art of Computer Programming for games.

Good luck.